Re: LT_AT_MAKE doesn't terminate the test on failure.

[Ralf Wildenhues]

* Peter Rosin wrote on Tue, Jun 08, 2010 at 03:07:05PM CEST:
> Sorry to reply to self, but the issue is that there is a subshell in
> LT_AT_MAKE (there's one in LT_AT_CONFIGURE too) which stops the
> exit from propagating all the way to where it's supposed to.

Good observation!  I really wonder how I managed to overlook that for so
long.

> Maybe adding the following after the end bracket of the sub-shells
> in those two macros is acceptable?
> 
> $at_failed && at_fn_log_failure

No, that is definitely too much internals for my taste.
This patch should work however.  OK?

Thanks,
Ralf

    Fix testsuite helper macros to not hide failure.
    
    * tests/testsuite.at (LT_AT_CONFIGURE, LT_AT_MAKE):
    Do not wrap AT_CHECK in a subshell, so that failures are
    properly propagated; instead, move environment changes inside
    the AT_CHECK code.
    Report and analysis by Peter Rosin.

diff --git a/tests/testsuite.at b/tests/testsuite.at
index ea092fc..da5463b 100644
--- a/tests/testsuite.at
+++ b/tests/testsuite.at
@@ -151,21 +151,19 @@ AT_KEYWORDS([autoconf automake])
 # LT_AT_CONFIGURE([OPTIONS])
 # --------------------------
 m4_define([LT_AT_CONFIGURE],
-[( SHELL=${CONFIG_SHELL-/bin/sh}
-   AT_CHECK([CONFIG_SHELL=$SHELL $SHELL ./configure $configure_options $1],
-        [0], [ignore], [ignore])
- )
+[AT_CHECK([: ${CONFIG_SHELL=/bin/sh}; export CONFIG_SHELL; ]dnl
+         [$CONFIG_SHELL ./configure $configure_options $1],
+         [0], [ignore], [ignore])
 ])
 
 
 # LT_AT_MAKE([TGTS], [VARS])
 # --------------------------
 m4_define([LT_AT_MAKE],
-[( $unset LIBTOOL LIBTOOLIZE
-  for target in m4_default([$1], [all])
-  do
-    AT_CHECK([$MAKE $target $2], [0], [ignore], [ignore])
-  done )
+[for target in m4_default([$1], [all])
+do
+  AT_CHECK([$unset LIBTOOL LIBTOOLIZE; $MAKE $target $2], [0], [ignore], 
[ignore])
+done
 ])